Abstrakt
Breast cancer characteristics obtained at the time of diagnosis are important for setting the basic strategy of the treatment. Reliability of preoperative investigation differs for various features of the disease.
The aim of this study was to ascertain the agreements and differences between preoperative and postoperative values. This retrospective study analyzed the results of 617 women with primary surgery of the breast and axilla.
Cohen's kappa coefficient has been employed to measure the degree of agreement between preoperative and postoperative values. Substantial or "almost perfect" agreement has been documented for the histological type of the tumors, their histopathological grade, proliferation index Ki67, as well as for estrogen, progesterone, and HER-2/neu receptors.
Substantial differences exist between preoperative and postoperative diagnoses of invasiveness of the tumor, determination of the size of the tumors, and the number of tumor foci. Preoperative imaging and clinical examination of lymph nodes exhibited unacceptably high false negative rates.
Heterogeneity of breast cancer cell population, methodology of histology examinations, and insufficient imaging of lymph nodes are the major limitations precluding satisfactory accuracy of preoperative diagnosis. Preoperatively diagnosed in situ carcinomas, as well as multifocal lesions, were the most often sources of diagnostic failures.